Strategy after AI adoption for higher productivity: exploration vs. exploitation

Abstract

As an increasing number of firms are adopting artificial intelligence (AI), understanding AI in the innovation strategic context is becoming more important. While the majority of previous literature has analyzed AI on an aggregate, national-level, and focused on its visible outcomes, further research is needed on strategies to employ at firm level in response to adopting AI. In this paper, we broaden insights on the studies of firms' innovation strategy after AI adoption by specifically focusing on firms that have adopted AI as part of services or products offered to customers. By extending exploration and exploitation literature in technological innovation context, we argue that impact on productivity differ depending on firms' subsequent exploration and exploitation focus after adopting AI. Productivity is measured using stochastic frontier analysis and each innovation direction is compared using meta-frontier analysis. In previous literatures, mixed views exist in the impact of AI on productivity. Moreover, prior research view the aggregate impact without considering much about firms’ behaviors after adopting AI. Thus, we prove that productivity improves when firms focus on exploitation after AI adoption, compared to when firms focus on exploration or no innovation. Our results shed light on the innovation strategy study and suggest innovation strategies for firms to pull with an emerging technology such as AI.